and Matthew MacManes - bioRxiv · 9 + [email protected] 10 ... 126 BUSCO searches against a database of highly-conserved single-copy genes in Arthropoda 127 ... of these conserved

Characterizing the Adult and Larval Transcriptome of the Multicolored Asian Lady Beetle, 1 Harmonia axyridis 2

3

Lindsay Havens 1* and Matthew MacManes 1+ 4

5

6

1. University of New Hampshire Department of Molecular Cellular and Biomedical Science 7

* [email protected] 8

+ [email protected] 9

10

11

12

Abstract: 13

The reasons for the evolution and maintenance of striking visual phenotypes are as widespread as the 14 species that display these phenotypes. While study systems such as Heliconius and Dendrobatidae 15 have been well characterized and provide critical information about the evolution of these traits, a 16 breadth of new study systems, in which the phenotype of interest can be easily manipulated and 17 quantified, are essential for gaining a more general understanding of these specific evolutionary 18 processes. One such model is the multicolored Asian lady beetle, Harmonia axyridis, which displays 19 significant elytral spot and color polymorphism. Using transcriptome data from two life stages, adult and 20 larva, we characterize the transcriptome, thereby laying a foundation for further analysis and 21 identification of the genes responsible for the continual maintenance of spot variation in H. axyridis. 22

23

24

25

26

27

28

29

30

31

.CC-BY 4.0 International licensenot certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (which wasthis version posted December 15, 2015. . https://doi.org/10.1101/034462doi: bioRxiv preprint

https://doi.org/10.1101/034462

http://creativecommons.org/licenses/by/4.0/

32

Introduction: 33

The evolution and maintenance of phenotypic polymorphism and striking visual phenotypes have 34 fascinated scientists for many years (Darwin, 1859; Endler, 1986; Fisher, 1930; Gray and McKinnon, 35 2006; Joron et al., 2006). In general, insects have become increasingly popular as study organisms to 36 examine phenotypic variation (Jennings, 2011; Joron et al., 2006). One such insect displaying extensive 37 elytra and spot variation that has yet to be extensively studied is the Asian Multicolored Ladybeetle, 38 Harmonia axyridis. 39

40

The mechanisms responsible for the evolution of these phenotypes are as widespread as the species 41 that display them. Aposematism, crypsis, and mimicry may play a role in the evolution of phenotypic 42 variation in the animal kingdom. Members of family Dendrobatidae, poison dart frogs, are 43 aposematically colored (Cadwell, 1996), while Tetrix subulata grasshoppers maintain their phenotypic 44 polymorphism to aid in crypsis (Karpestam et al., 2014). A mimicry strategy is utilized by one particularly 45 well-characterized species that exhibits phenotypic polymorphism, the Neotropic butterfly system, 46 Heliconius. The color, pattern, and eyespot polymorphism seen in Heliconius is thought to have arose 47 as a result of Müllerian mimicry (Flanagan et al., 2004) and the supergenes underlying these traits have 48 been well characterized (Kronforst et al., 2006, Joron et al., 2006, Jones et al., 2012). 49

50

These studies, aiming to elucidate the mechanistic links between phenotype and genotype, present a 51 unique opportunity to gain insight into the inner workings of many important evolutionary processes. 52 While systems like poison frogs and butterflies have been pioneering, the use of novel models, 53 especially those that can be easily manipulated, are needed. One such study system that possesses 54 many of the benefits of classical models, while offering several key benefits is the multicolored Asian 55 lady beetle, Harmonia axyridis. Harmonia, which is common throughout North America, and easily bred 56 in laboratory environments, possesses significant variation in elytral spot number and color. 57

58

Elytra color can be red, orange, yellow, or black and spot numbers of H. axyridis range from zero to 59 twenty-two (personal observation). The patterning is symmetrical on both wings. In some animals, there 60 is a center spot beneath the pronotum which leads to an odd number of spots. The elytral spots are 61 formed by the production of melanin pigments (Bezzerides et al., 2007). The frequency of different 62 morphs varies with location and temperature. The melanic morph is more prevalent in Asia when 63 compared to North America (LaMana and Miller, 1996; Dobzhansky, 1933). A decrease in melanic H. 64 axyridis has been shown to be correlated with an increase in average yearly temperatures in the 65 Netherlands (Brakefield and de Jong, 2011). 66


https://doi.org/10.1101/034462


67

Sexual selection may play a role in color variation in H. axyridis. Osawa and Nishida (1992) remarked 68 that female H. axyridis might choose their mates based on melanin concentration. Their choice, 69 however, has been shown to vary based on season and temperature. Non-melanic (red, orange, or 70 yellow with any spot number) males have a higher frequency of mating in the spring-time, while melanic 71 (black) males have an increased frequency of mating in the summer. While this has been shown with 72 respect to elytral color, no such findings have occurred for spot number. Although these spot patterns 73 are believed to be related to predator avoidance, thermotolerance, or mate choice (Osawa and Nishida, 74 1992), the genetics underlying these patterns is currently unknown. 75

76

To begin to understand the genomics of elytral coloration and spot patterning, we sequenced the 77 transcriptome of an late-stage larva and adult ladybug. These results lay the groundwork for future study 78 of the genomic architecture of pigment placement and development in H. axyridis. 79

80

Methods and Materials: 81

Specimen capture, RNA extraction, library prep and sequencing 82

83

One larval (Figure 1a) and one adult (Figure 1b) H. axyridis were captured on the University of New 84 Hampshire campus in Durham, New Hampshire (43.1339° N, 70.9264° W). The adult was orange with 85 18 spots. The insects were immediately placed in RNAlater and stored in a -80C freezer until RNA 86 extraction was performed. The RNA from both individuals was extracted following the TRIzol extraction 87 protocol (Invitrogen, Carlsbad USA). The entire insect was used for the RNA extraction protocol. The 88 quantity and quality of extracted RNA was analyzed using a Qubit (Life Technology, Carlsbad USA) as 89 well as a Tapestation 2200 (Agilent technologies, Palo Alto USA) prior to library construction. Following 90 verification, RNA libraries were constructed for both samples following the TruSeq stranded RNA prep 91 kit (Illumina, San Diego USA). To allow multiple samples to be run in one lane, a unique adapter was 92 added to each sample. These samples were then pooled in equimolar quantities. The libraries were 93 then sent to the New York Genome Center (New York, USA) for sequencing on a single lane (125bp 94 paired end) of the HiSeq 2500 sequencer. 95

96


https://doi.org/10.1101/034462


97 98 Figure 1a: The larva used for transcriptome sequencing. 99 100

101 102 Figure 1b: The adult used for transcriptome sequencing. 103 104 105 Sequence Data Preprocessing and Assembly 106 107 The raw sequence reads corresponding to the two tissue types were error corrected using the software 108 BLESS (Heo et al., 2014) version 0.17 (https://goo.gl/YHxlzI, https://goo.gl/vBh7Pg). The error-109 corrected sequence reads were adapter and quality trimmed following recommendations from 110 MacManes (MacManes, 2014) and Mbandi (Mbandi, 2014). Specifically, adapter sequence 111 contamination and low quality nucleotides (defined as Phred<2) were removed using the program 112


https://doi.org/10.1101/034462


Trimmomatic version 0.32 (Bolger, 2014) called from within the Trinity assembler version 2.1.1 (Haas, 113 2013). Reads from each tissue were assembled together to created a joint assembly of adult and larva 114 transcripts using a Linux workstation with 64 cores and 1Tb RAM. We used flags to indicate the stranded 115 nature of sequencing reads and set the maximum allowable physical distance between read pairs to 116 999nt (https://goo.gl/ZYP08M). 117 118 The quality of the assembly was evaluated using transrate version 1.01 (Unna-Smith 2015; 119 (https://goo.gl/RpdQSU). Transrate generates quality statistics based on a process involving mapping 120 sequence reads back to the assembled transcripts. Transcripts supported by properly mapped reads of 121 a sufficient depth (amongst other things) are judged to be of high quality. In addition to generating quality 122 metrics, transrate produces an alternative assembly with poorly-supported transcripts removed. This 123 improved assembly was used for all downstream analyses and QC procedures. We then evaluated 124 transcriptome completeness via use of the software package BUSCO version 1.1b (Simão 2015). 125 BUSCO searches against a database of highly-conserved single-copy genes in Arthropoda 126 (https://goo.gl/bhTNdr). High quality, complete transcriptomes contain are hypothesized to contain the 127 vast majority of these conserved genes. 128 129 To remove assembly artifacts remaining after transrate optimization, we estimated transcript abundance 130 using 2 software packages - Salmon version 0.51 (Patro 2015; https://goo.gl/01UIF6) and Kallisto 131 version 0.42.4 (Bray, 2015; https://goo.gl/BsQMpr). Transcripts whose abundance exceeded 0.5 TPM 132 in either adult or larval datasets using either estimation method were retained. We evaluated 133 transcriptome completeness and quality, again, after TPM filtration, using BUSCO and transrate, to 134 ensure that our filtration processes did not significantly effect the biological content of the assembly. 135

136

Assembled Sequence Annotation 137 138 The filtered assemblies were annotated using the software package dammit 139 (https://github.com/camillescott/dammit; https://goo.gl/05MY5i). Dammit coordinates the annotation 140 process, which involves use of blast (Camacho, 2009), TransDecoder (version 2.0.1, 141 http://transdecoder.github.io/), and hmmer version 3.1b1 (Wheeler 2013). In addition to this, putative 142 secretory proteins were identified using the software signalP, version 4.1c (https://goo.gl/FaOQSj). 143 144 To identify patterns of gene expression unique to each life stage, we used the expression data as per 145 above. We identified transcripts expressed in one stage but not the other, and cases where expression 146 occurred in both life stages. The Uniprot ID was identified for each of these transcripts using a blastx 147 search (https://goo.gl/J9saMj), and these terms were used in the web interface Amigo (Carbon et al., 148 2009) to identify Gene Ontology terms that were enriched in either adult or larva relative to the 149


https://doi.org/10.1101/034462


background patterns of expression. The number of unique genes contained in the joint assembly was 150 calculated via a BLAST search of the complete gene sets of Homo sapiens, Drosophila melanogaster, 151 and Tribolium casteneda. 152

153

Results and Discussion: 154

155

Data Availability 156 157 All read data are available via https://goo.gl/VNuHXC. Assemblies and data matrices are available at 158 https://goo.gl/D3xh65. 159

160

RNA extraction, Assembly and Evaluation 161

RNA was extracted from whole bodies of both the adult and the larva stage of a single Harmonia 162 axyridis. The quality was verified using a Tapestation 2200 as well as a Qubit. The initial concentration 163 for the larva sample was 83.2 ng/uL, while the initial concentration for the adult sample was 74.7 ng/uL. 164 The number of strand-specific paired end reads contained in the adult and larva libraries were 58 million 165 and 67 million, respectively. The reads were 125 base pairs in length. 166

167

The raw Trinity assembly of the larval and adult reads resulted in a total of 171,117 contigs (82Mb) 168 exceeding 200nt in length. This assembly was evaluated using Transrate, producing an initial score of 169 0.10543, and and optimized score of 0.29729. This transrate optimized assembly (89,305 transcripts, 170 62Mb) was further filtered by removing transcripts whose expression was less than 0.5 TPM. After 171 filtration, 33,648 transcripts (40Mb) remained. To assess for the inadvertent loss of valid transcripts, we 172 ran BUSCO before and after this filtration procedure. The percent of Arthropoda BUSCO’s missing from 173 the assembly rose slightly, from 18% to 21%. Transrate was run once again, and resulted in a final 174 assembly score of 0.29112. This score is indicative of a high-quality transcriptome appropriate for 175 further study. In an attempt at understanding how many distinct genes our transcriptome contained, we 176 conducted a blast search against Homo sapiens, Drosophila melanogaster, and Tribolium casteneda. 177 This search resulted in 7,246, 7,739, and 7,741 unique matches, which serve as estimates of the 178 number of unique genes expressed in these two life stages. The final assembly is available at 179 https://goo.gl/nWdBuv 180

181

Annotation 182

183


https://doi.org/10.1101/034462


The assembled transcripts were annotated using the software package dammit!, which provided 184 annotations for 23,304, or 69% of the transcripts (available here: https://goo.gl/gpGXLG). These 185 annotations included putative protein and nucleotide matches, 5- and 3-prime UTRs, as well as start 186 and stop codons. In addition to this, analysis with Transdecoder yielded 14,518 putative protein 187 sequences (available here: https://goo.gl/qVLWwD), which were annotated by 4,139 distinct Pfam 188 protein families, while 176 transcripts were determined to be non coding (ncRNA) based on significant 189 matches to the Rfam database (available here: https://goo.gl/x1n7jC). Lastly, 2,925 proteins (7.8% of 190 total) were determined to to be secretory in nature by the software package signalP (available here: 191 https://goo.gl/z0ra1g). 192

193

Annotation of the sequence dataset resulted in the identification of host of transcripts that may be of 194 interest to other researchers including: 43 heat-shock and 8 cold-shock transcripts, 87 homoebox-195 domain containing transcripts, 122 7-transmembrane-containing (18 GPCR’s) transcripts, 13 solute 196 carriers, 143 ABC-transport-containing transcripts, and 21 OD-S (pheromone-binding) transcripts. 197

198

A complement of immune-related genes were discovered as well. These include a single member of 199 the Attacins and Coleoptericins, two TLR-like genes, seven Group 1, and 34 Group 2 C-type lectin 200 Receptors (CLRs). Two CARD-containing Cytoplasmic pattern recognition receptor (CRR) genes were 201 discovered, as were 3 MAP kinase containing transcripts. Finally, 119 RIG-I-like receptors (RLR) were 202 found. 203

204

Gene expression was estimated for each transcript for both adult and larva (Figure 2, available here: 205 https://goo.gl/wM3TV7). For all transcripts expressed in the adult, the mean TMP=29.7 (max= 80,016.9, 206 SD= 500.8) while the mean larval TPM=29.71 (max= 166,264, SD=941). When analyzing transcripts 207 found uniquely in these two tissues, mean adult TMP=9.9 (max=3,037, SD=86) and for larva TPM=8.6 208 (max=687, SD=41). 209


https://doi.org/10.1101/034462


210

211

Figure 2: The Venn diagram representing the number of transcripts expressed in both adult and 212 larva, as well as those expressed uniquely in one or the other. 213

214

Analysis of the differences between adult and larval life stages were carried out as well. Because these 215 life stages were only sequenced with a single individual each, they should be interpreted with some 216 caution. The vast majority of transcripts were observed in both life stages (n=30,630, 91%), with a small 217 number being expressed uniquely in larva (n=1,094) and adult (n=1,922). Of these transcripts 218 expressed uniquely in either larva or adult, 6.1% and 4.6%, respectively, were found to be secretory in 219 nature. 220

221

Conclusions: 222

Phenotypic polymorphisms and striking visual phenotypes have fascinated scientists for many years. 223 The breadth of evolutionary causes for the maintenance of these phenotypes are as numerous as the 224 species that display them. One organism, Harmonia axyridis, provides a unique opportunity to explore 225 the genetic basis behind the maintenance of an easy to quantify variation - elytral spot number. While 226 understanding these genomic mechanisms is beyond the scope of this paper, we do provide a reference 227 transcriptome for H. axyridis, a foundational resource for this work. 228

229

This study indicates that most gene expression profiles are shared across life stages of H. axyridis. 230 While the majority of proteins identified in the assembled transcriptome were structural in function, 231 analyses of protein families using the Pfam database indicated the presence of pigment proteins. In 232

30630transcripts expressed in both

adult and larva

1922 transcripts expressed in adult only

1094 transcripts expressed in larvaonly


https://doi.org/10.1101/034462


particular, RPE65, which functions in the cleavage of carotenoids, was found. In H. axyridis, increased 233 carotenoid pigmentation has been linked to increased alkaloid amounts (Britton et al., 2008). In addition, 234 the elytral coloration of the seven spot ladybug, Coccinella septempunctata, is a result of several 235 carotenoids (Britton et al., 2008). While larva are mostly black (Figure 1a), we posit that the orange 236 sections on the lower back could be due to carotenoid production. Moreover, this study provides a 237 necessary foundation for the continued study of the genetic link between genes and the maintenance 238 of variation in H. axyridis. 239

240 241


https://doi.org/10.1101/034462


Literature Cited: 242

Bezzerides A. L., McGraw K. J., Parker R. S., Husseini J. (2007). Elytral color as a signal of chemical 243 defense in the Asian ladybird beetle Harmonia axyridis. Behavior Ecology and Sociobiology 61: 244 1401-1408. 245

Bolger A. M., Lohse M., Usadel., B. (2014). Trimmomatic: A flexible trimmer for Illumina Sequence Data. 246 Bioinformatics, btu170. 247

Brakefield P. M., de Jong P. W. (2011). A steep cline in ladybird melanism has decayed over 25 248 years: a genetic response to climate change? Heredity 107: 574-578. 249

Bray N., Pimental H., Melsted P., Lior P. (2015). Near-optimal RNA-seq quantification. eprint 250 arXiv:1505.02710 251

Britton G., Liassen-Jesen, S., Pfander H.(2008). Carotenoids, Vol. 4: Natural Functions. Springer. DIO: 252 0.1007/978-3-7643-7499-0 253

Cadwell J. P., (1996). The Evolution of myrmecophagy and its correlates in poison frogs (Family 254 Dendrobatidae). Journal of zoology 240: 75-101. 255

Camacho C., Coulouris G., Avagyan V., Ma N., Papadopoulos J., Bealer K., Madden T. L. (2009). 256 BLAST+: architecture and applications. BMC Bioinformatics doi: 10.1186/1471-2105-10-421. 257

Carbon S., Ireland A., Mungall C.J., Shu S., Marshall B., Lewis S. (2009). AmiGO Hub, Web Presence 258 Working Group. AmiGO: online access to ontology and annotation data. Bioinformatics 259 25(2):288-9. 260

Darwin C. (1859). The Origin of Species. New York, New York: Bantam Dell. 261

Dobzhansky T. (1993). Geographical variation in ladybeetles. The American Naturalist 67: 97-126. 262

Endler J. A. (1986). Natural Selection in the Wild. Princeton, New Jersey: Princeton University Press. 263

Fisher, R. A. (1930). The Genetical Theory of Natural Selection. Oxford: Clarendon Press. 264

Flanagan N.S., Tobler A., Davison A., Pybus O.G., Kapan D. D., Planas S., Linares M., Heckel D., 265 McMillan W. O. (2004). Historical demography of Müllerian mimicry in the neotropical Heliconius 266 butterflies. PNAS 101: 9704-9709. 267

Gray S. M., McKinnon J. S. (2006). Linking color polymorphism maintenance and speciation. 268 ScienceDirect: doi:10.1016/j.tree.2006.10.005 269

Haas B. J., Papanicolaou A., Yassour M., Grabherr M., Blood P. D., Bowden J., Couger M. B., Eccles 270 D., Li B., Lieber M., MacManes M. D., Ott M., Orvis J., Pochet N., Strozzi F., Weeks N., 271 Wasterman R., William T., Dewey C. N., Henschel R., LeDuc R. D., Friedman N., Regev A. 272 (2013). De novo transcript sequence reconstruction from RNA-seq using Trinity platform for 273 reference generation an analysis. Nature Protocols 8: 1494-1512. 274


https://doi.org/10.1101/034462


Heo Y., Wu X., Chen D., Ma J., H W. (2013). BLESS: Bloom filtered-based error correction solution for 275 high-throughput sequencing reads. Bioinformatics doi: 10.1093/bioinformatics/btu030. 276

Jennings B. H. (2011). Drosophila – a versatile model in biology & medicine. MaterialsToday 14: 190-277 195. 278

Jones R.T., Salazar P. A., ffrench-Constant R. H., Jiggins C.D., Joron M. (2012). Evolution of a mimicry 279 supergene from a multilocus architecture. Proc. R. Soc. B 279: doi: 10.1098/rspb.2011.0882. 280

Joron M., Jiggins C. D., Papanicolaou A., McMillan W. O. (2006). Heliconius wing patterns: an evo-281 devo model for understanding phenotypic diversity. Heredity 97: 157-167. 282

Karpestam E., Merilaita S., Forsman A. (2014). Natural levels of colour polymorphism reduce 283 performance of visual predators searching for camouflaged prey. Biol J Linnean Soc. 284 doi: 10.1111/bij.12276. 285

Kronforst M. R., Young L. G., Kapan D. D., McNeely C., O’Neill R. J., Gilbert L. E. (2006). Linkage of 286 butterfly mate preference and wing color preference cue at the genomic location of wingless. 287 PNAS 103: 6575-6580. 288

LaMana M.L., Miller J.C. (1996). Field observations on Harmonia axyridis Pallas (Coleoptera: 289 Coccinellidae) in Oregon. Biological Control 6: 232-237. 290

MacManes M. D. (2014). On the optimal trimming of high-throughput mRNA sequence data. Front 291 Genet 5: 13. doi: 10.3389/fgene.2014.00013. 292

Mbandi S. K., Hesse U., Rees J. G., Christoffels A. (2014). A glance at quality score: implication for 293 de novo transcriptome reconstruction of Illumina reads. Frontiers in Genetics (5): doi: 294 10.3389/fgene.2014.00017 295

Osawa N., Nishida T., (1992). Seasonal variation in elytral colour polymorphism in Harmonia axyridis 296 (the ladybird beetle): the role of non-random mating. Heredity 69: 297-307. 297

Patro R., Suggal G.,Kingsford C. (2015). Accurate, fast, and model-aware transcript expression 298 quantification with Salmon. bioRxiv: doi: http://dx.doi.org/10.1101/021592 299

Simão F. A., Waterhouse R. M., Ioannidis P., Kriventseva E. V., Zdobnov E. M. (2015). BUSCO: 300 assessing genome assembly and annotation completeness with single-copy orthologs. 301 Bioinformatics dio: 10.1093/bioinformatics/btv351. 302

Smith-Unna R. D., Boursnell C., Patro R., Hibberd J. M., Kelly S. (2015). TransRate: reference free 303 quality assessment of de-novo transcriptome assemblies. bioRxiv: dio 304 http://dx.doi.org/10.1101/021626. 305

Wheeler T. J., Eddy S. R. (2013). nhmmer:DNA homology search with profile HMMs. (2013). 306 Bioinformatics: doi: 10.1093/bioinformatics/btt403 307


https://doi.org/10.1101/034462


Documents

and Matthew MacManes - bioRxiv · 9 + [email protected] 10 ... 126 BUSCO searches against a database of highly-conserved single-copy genes in Arthropoda 127 ... of these conserved